organic andaqueous phases (Whittemore etal., 1969;. Gardas & Koscielak, 1971; Liotta et al., 1972; Rosse. & Lauf, 1970). Blood-group-A and blood-group-B.
Biochem. J. (1974) 138, 381-386 Printed in Great Britain
381
The Distribution of Blood-Group Antigens on Butanol Extraction of Human Erythrocyte 'Ghosts' By D. J. ANSTEE and M. J. A. TANNER South West Regional Blood Transfusion Centre, Southmead, Bristol BS10 5ND, U.K., and Department of Biochemistry, University ofBristol, Bristol BS8 1 TD, U.K. (Received 4 September 1973)
The distribution of protein and blood-group-antigen activity obtained after butanol extraction oferythrocyte 'ghosts' under various conditions is described. Butanol extraction under low-ionic strength conditions results in the recovery of membrane protein in high yield in the aqueous phase. Blood-group-A activity is found in both the aqueous and butanol phases, whereas blood-group-P activity is confined to the butanol phase and blood-group-I and blood-group-MN activity are restricted to the aqueous phase. Much lower yields of protein are obtained in the aqueous phase when high-ionic-strength conditions are used. An appreciable amount of material is precipitated at the interface. Under these conditions blood-group-P activity is found only in the butanol phase, blood group-A activity in the butanol phase and interface material and only blood-groupMN activity in the aqueous phase. In contrast with previous reports no correlation could be demonstrated between the secretor status of the donors and the presence of blood-groupA activity in the aqueous phase after butanol extraction under any of the extraction conditions used. By using butanol extraction under high-ionic-strength conditions it is possible to isolate the blood-group-MN-active sialoglycoprotein in high yield from erythrocyte 'ghosts' by a simple procedure.
The butanol-extraction technique (Maddy, 1966) has been widely used to effect a gross separation of erythrocyte 'ghost' proteins from erythrocyte lipids. There have been several reports of the application of this technique to study the distribution of human erythrocyte blood-group antigens between the organic and aqueous phases (Whittemore et al., 1969; Gardas & Koscielak, 1971; Liotta et al., 1972; Rosse & Lauf, 1970). Blood-group-A and blood-group-B activities have been found in both phases and this observation has led to the proposition that two distinct types of A and B antigen exist. These two forms of A and B antigen are assumed to reflect differences in the nature of the molecules that carry the antigenic determinants. Gardas & Koscielak (1971) reported that the presence of the secretor gene correlated with the appearance of blood-groupA and blood-group-B activity in the aqueous phase obtained after butanol extraction of erythrocyte 'ghosts'. We report here our observations on the distribution of membrane proteins and bloodgroup-antigen activity under various conditions of butanol extraction of human erythrocyte 'ghosts'. Materials and Methods General methods Protein was determined by the method of Lowry et al. (1951) and sialic acid by the thiobarbituric Vol. 138
acid method (Warren, 1959). Polyacrylamide-gel electrophoresis in the presence of 0.1% sodium dodecyl sulphate was carried out by the method of Fairbanks et al. (1971). The preparation of samples and staining of the gels was as described by Tanner & Boxer (1972). Samples were taken immediately after preparation, treated as described above and stored frozen before electrophoresis.
Erythrocyte 'ghosts' Samples (10 or 20ml) offresh blood were drawn from volunteers with the use of acid-citrate-dextrose anticoagulant and used within 1 week. The cells were then washed three times with 0.15 M-NaCl. All preparations of erythrocyte 'ghosts' were stored at 40C. To prepare 'Dodge ghosts' (Dodge et al., 1963) the cells were washed twice with 50ml of 0.103MNa2HPO4 adjusted to pH 7.4 with 0.155 M-NaH2PO4 (iso-osmotic phosphate buffer), and lysed by dilution into 200ml of cold iso-osmotic phosphate buffer previously diluted with 20.5 vol. of cold de-ionized water. The 'ghosts' were collected by centrifugation at 20000g for 40min and washed once with the same solution to yield a readily dispersed pale yellowishpink 'ghost' preparation. 'Koscielak ghosts' were prepared as described by Gardas & Koscielak (1971). The washed cells were lysed into 5 vol. of cold de-ionized water (which had been titrated to pH 5.5 with 1 M-acetic acid) and kept at 4°C for 30min.
382 The 'ghosts' were collected by centrifugation at 20000g for 5min and repeatedly washed in the centrifuge with the lysing solution until the supernatant was colourless. A red, rather sticky pellet was obtained. Phase-contrast microscopy showed the 'Dodge ghosts' to be freely mobile, discoid, intact erythrocyte 'ghosts', whereas the 'Koscielak ghosts' preparations consisted of highly aggregated small vesicles.
Butanol extraction This was done at 4°C on freshly prepared 'ghosts'. The 'ghosts', derived from 10ml of whole blood, were suspended in 50ml of the required buffer, centrifuged, and the pellet was made up to 5.5 ml with the same buffer. Part of this suspension (4.5 ml) was used for butanol extraction by the method of Gardas & Koscielak (1971). Ice-cold n-butanol (2.25 ml) was added to each sample, the mixture was vigorously shaken for 20s and then kept on ice for 15 min. The phases were separated by centrifugation at 40000g for 15min. The aqueous phase was carefully removed and 2.5ml of the aqueous buffer with which the 'ghosts' were originally equilibrated was added to the butanol phase and interface material. The mixture was shaken vigorously for 20s, centrifuged at 40000g for 10min and the aqueous phase was pooled with the previously obtained aqueous phase. Cold butan-1-ol (0.5vol.) was added to the pooled aqueous phases, the mixture was shaken for 20s and then kept for 5min on ice. The aqueous phase was removed and the butanol phase and any interface was pooled with the previously obtained butanol phase and interface. The pooled butanol phases and interfaces were dried in vacuo. Ethanol (83 %, v/v; 3.5 ml) was added to the dry material and the mixture was heated in boiling water for 3 min. The ethanolsoluble material was decanted into a pointed centrifuge tube and left overnight at -20°C. The precipitate was collected by centrifugation at -10°C for 30min at 20000g, resuspended in 5ml of acetone at -20°C and the centrifugation repeated. The final washed precipitate was dried under a stream of N2. The aqueous phase was dialysed overnight at 4°C against two changes of 2 litres of phosphate-buffered saline (0.15M-NaCl in 10mM-phosphate buffer, pH7.0). The dialysed solutions were concentrated to a volume of 2.5-3.5 ml (for every 10ml of whole blood used in the preparation) at 4°C by coating the dialysis bags with dry Sephadex G-100 and were stored frozen. These solutions were assayed for antigen activity as soon as possible after their preparation (usuallywithin 2-3 days). Preparation oferythrocyte sialoglycoprotein 'Dodge ghosts' were prepared from 200ml of packed outdated erythrocytes. The final 'ghosts'
D. J. ANSTEE AND M. J. A. TANNER were washed with 20mM-sodium phosphate buffer, pH 7.0, and made up to 180ml with the same buffer. Butanol extraction of these 'ghosts' was as described above. The aqueous phase was dialysed overnight at 4°C against two changes of 8 litres of 0.1mMsodium phosphate buffer, pH7.0. This solution was concentrated by pressure dialysis under N2 over an Amicon PM-10 membrane to 20ml and the sample was applied to a column (2.5 cm x 80 cm) of Sepharose 6B, previously equilibrated with a solution containing 50mM-Tris-HCl, pH8.0, 0.1mm-EDTA and 0.2MNaCl. The relevant fractions were pooled, dialysed exhaustively against water at 4°C and freeze-dried to yield 53mg dry weight of the sialoglycoprotein
preparation. Serological methods Human anti-A, anti-B, rabbit anti-M and anti-N antisera were from the Blood Group Reference Laboratory, London, U.K. Human anti-I (fry) antiserum was the gift of Miss M. G. Kenwright, London, U.K. Goat anti-P1 antiserum was kindly given by Professor 0. Prokop, Berlin, G.D.R. and anti-P (ellis) antiserum was from Dr. J. Darnborough, Cambridge, U.K. Four auto-anti-I sera were used. All other antisera were available from the Regional Blood Transfusion Centre, Bristol, U.K. Cell typing was performed by the standard serological techniques appropriate for the particular antiserum used [saline, albumin layering, indirect anti-human globulin methods (Boorman & Dodd, 1970)]. Manual haemagglutination-inhibition tests were performed as described by Anstee et al. (1973). After sonication of interface fractions [in 1 ml of phosphate-buffered saline (Dulbecco 'A'; Oxoid Ltd., London E.C.4, U.K.) pH7.3], much insoluble material remained and the supernatants only were examined for blood-group activity after centrifugation at 900g for 10min at 4°C. Manual haemagglutination-inhibition assessments are not satisfactory for the calculation of yields of blood-group activity in the separated phases because of the inherent inaccuracies of the system. A quantitative system of haemagglutination inhibition was therefore preferred. Manual assessments gave an essentially similar distribution of blood-group activity to that obtained by quantitative assessment. Quantitative haemagglutination-inhibition assessments were done by using the method of Gibbs et al. (1961), modified for use in a Technicon Autoanalyser system. The agglutination obtained with various dilutions of antibody is calculated as a percentage of the theoretical maximum agglutination. A plot of percentage agglutination against antibody dilution on probit/treble cycle log paper gives a value for the amount of antibody necessary to give 50% agglutination in the system (1 HD50 unit); 1974
383
BUTANOL EXTRACTION OF HUMAN ERYTHROCYTE 'GHOSTS'
5HD50 units of antibody are then sampled at the same time as various concentrations of inhibitor. A plot of percentage agglutination in the presence of inhibitor against inhibitor concentration gives a value for HID50 (which is the amount of inhibitor that is equivalent to 4HD5o units of antibody). Results Distribution ofantigens
Each of the butanol-extraction experiments (Table 2) were carried out on six individual donations of blood. In each experiment three of the donations were from blood-group-Al secretors and three were from blood-group-A1 non-secretors (Table 1). Three separate extraction experiments were carried out with butanol (Table 2). In Expts. 1 and 2 intact 'ghosts' were prepared by the procedure of Dodge et al. (1963) ('Dodge ghosts') and then equilibrated either with lmM-sodium phosphate buffer, pH 7.0 (Expt. 1) or with 20mM-sodium phosphate buffer, pH7.0 (Expt. 2) before butanol extraction by the method of Gardas & Koscielak (1971). In Expt. 3, 'ghosts' ('Koscielak ghosts') were prepared by the method of Gardas & Koscielak (1971), equilibrated with 20mM-sodium phosphate buffer, pH7.0 and extracted as described by these authors. The recoveries of protein in the aqueous phase were 65 (±4)%, 4 (±1)% and 32 (±5)% in Expts. 1, 2 and 3 respectively. When the butanol extraction was done on 'Dodge ghosts' at very low ionic strengths then the recovery of protein in the aqueous phase was quite high. In contrast, the presence of 20mM buffer in the aqueous medium dramatically lowered the recovery of protein in the aqueous phase. An appreciable amount of material was found precipitated at the interface in this latter situation. Sodium dodecyl sulphate-polyacrylamide-gel electrophoretograms (Plate lb) show that the aqueous phase of Expt. 1
yielded a protein pattern which confirms that all the proteins that are present in the intact erythrocyte 'ghosts' are retained in the aqueous phase although it is depleted in protein E. In contrast, the aqueous phase derived from Expt. 2 yields essentially only one weak band on staining for protein which corresponds in mobility to the erythrocyte sialoglycoprotein (Plate ld). This band also stains with the periodic acid-Schiff's reagent and is the erythrocyte sialoglycoprotein (Plate le). The 'ghosts' obtained by the 'Koscielak procedure' gave variable gel-electrophoresis patterns which appeared to show evidence of protein degradation (Plate if), were grossly contaminated with haemoglobin and had no morphological resemblance to the 'Dodge ghosts'. Although there is an apparently high recovery of protein in the aqueous phase obtained on butanol extraction of these 'ghosts', gel electrophoresis showed that much of this is haemoglobin, and few other discrete bands that could be ascribed to 'ghost' proteins were found. The interface in this case also contained a large amount of protein (Plate 1h). The distribution of blood-group-A activity in the original 'ghosts', and in the fractions obtained after extraction, was estimated by automated quantitative haemagglutination inhibition and the results are shown in Table 2. In Expt. 1 a good overall recovery of blood-group-A activity was obtained. Most of this activity was found in the aqueous phase, only about 10% being in the butanol phase. In contrast, identical 'ghosts' extracted in the presence of 20mM-sodium phosphate buffer (Expt. 2) showed no recovery of blood-group-A activity in the aqueous phase, whereas the butanol phase contained significantly more blood-group-A activity than did similar material derived from Expt. 1. Under these conditions a good overall recovery was again obtained, appreciable amounts of blood-group-A activity being found in the interface material. 'Koscielak ghosts' extracted under the latter conditions (Expt. 3)
Table 1. Antigenic profile of intact erythrocytes from donors used in experiments Cell and saliva typing were done by using standard serological techniques. N.t. denotes 'not tested'. Secretor Antigen M P1 status Donors B Leb Saliva* Lea Al _ _ + + + ~~+ + Secretors U.C. + + + R. B. + + + + + P. H.t + Non-secretors A. W. -+___ + R. C. NA + + N.t. N.t. A. F. + + N.t. N.t. J. E. + + N.t. N.t. + + H. D. * + or - denotes presence or absence of blood-group-A activity in saliva samples.
t Weak A1 reactor. Vol. 138
N _
_
+
_
+
N.t.
N.t. N.t.
N.t. N.t.
N.t.
Q_
384
,0X
0 .4)
oo't
D. J. ANSTEE AND M. J. A. TANNER
eqOe
e
r- Ioo oo Vu400
en
- OIn oI0eq OC oe V" d-te
en_ N 4 q (In
-
0
''0
CU
E
